GPUs Over People? How VCs are Funding the AI Future

02:57:53.100 –> 02:57:59.900
Julia Nimchinski: And now we turn to the venture lens with ken fine CEO of affinity.

1008
02:58:00.360 –> 02:58:05.730
Julia Nimchinski: and really excited to explore where the capital is flowing. Welcome to the show ken.

1009
02:58:05.950 –> 02:58:07.759
Ken Fine: Thank you. Great to be here.

1010
02:58:08.430 –> 02:58:11.330
Julia Nimchinski: Running any triathlons lately.

1011
02:58:11.920 –> 02:58:19.642
Ken Fine: I need triathlons. I wasn’t expecting that topic. My next triathlon is in the fall, and I’m starting to train

1012
02:58:20.930 –> 02:58:26.080
Ken Fine: so that that could be a that could be a fun session into itself. Are we ready to kick off here.

1013
02:58:26.280 –> 02:58:27.979
Julia Nimchinski: Yep! Stage your swords.

1014
02:58:28.200 –> 02:58:48.020
Ken Fine: Okay, then we shall do so. Why don’t I begin with introductions? So I’m moderating the Vc roundtable today where we’ll be looking at the impact of AI on Vc investing. I’m the CEO of affinity. I’ve been at affinity for about a year, spent the previous 2527 years in high growth. B, 2 B Saas companies, multiple different roles

1015
02:58:48.140 –> 02:58:50.030
Ken Fine: running product marketing.

1016
02:58:50.230 –> 02:58:52.130
Ken Fine: Cco, CEO

1017
02:58:52.700 –> 02:59:12.549
Ken Fine: have seen b 2 b. Saas in many different stages and many different perspectives. Why don’t we do some introductions? And then we can talk about the topics. I’ll give a little background about affinity, and then hand off to the panel, and then I’ll come back and tee things up. Affinity is a relationships and relationship intelligence platform

1018
02:59:12.650 –> 02:59:31.970
Ken Fine: used in private capital by venture capitalists and private equity as well. Essentially, we enable our customers to leverage their full relationship, graph or network in service of workflows, such as sourcing deal, flow, portfolio management, other front office activities

1019
02:59:32.480 –> 02:59:35.689
Ken Fine: one. I hand off to the panel to introduce themselves.

1020
02:59:36.330 –> 02:59:37.780
Ken Fine: Jeremy, should we start with you.

1021
02:59:38.650 –> 02:59:53.220
jeremyk: Sure. Hey, everybody! I’m Jeremy Kaufman. I’m a partner at scale venture partners in San Francisco focus on series A and series B Enterprise tech companies, and I personally spend a lot of time at the application layer. So excited to chat with everybody.

1022
02:59:53.880 –> 02:59:55.540
Ken Fine: Excellent Thomas.

1023
02:59:57.050 –> 03:00:12.519
Thomas Cuvelier: And so, Thomas, I’m a partner at Rtp global, we’re an early stage fund focusing on on seed and series a globally across us, Europe and India, and I spend most of my time on the application layer as well.

1024
03:00:13.170 –> 03:00:15.209
Ken Fine: Very good welcome, and Kathleen.

1025
03:00:16.420 –> 03:00:32.479
Kathleen Estreich: Hi, I’m Kathleen. I’m a partner at Pair, a proud supporter of affinity who’s a pair portfolio company. So I we focus on early stage. So pre seed and seed and I focus mostly on the application layer as well.

1026
03:00:33.110 –> 03:00:36.540
Ken Fine: And thank you, Kathleen, for your support, and and finally.

1027
03:00:37.370 –> 03:01:03.970
Hadley: Hi, my name is Hadley Harris, co-founder and general partner of Eniac ventures. We’re also infinity user. Yeah, I I mainly focus on AI both kind of application layer and more kind of tooling. Lower in the stack. Been working in AI a long time, built a company called Lingo, those 1st voice based assistant and later kind of got folded into Siri back in 2,007 to 2,011. So excited to be here.

1028
03:01:04.340 –> 03:01:13.330
Ken Fine: Excellent great to have all of you here. I’ll tee up the topics and then we’ll get started. So again, we’re looking at the impact of AI and Vc investing. And we’ll also talk about some

1029
03:01:13.690 –> 03:01:18.839
Ken Fine: trends, input suggestions for founders. CEO executives in the space

1030
03:01:19.534 –> 03:01:28.430
Ken Fine: some of the subtopics we’re about to cover include how agentic AI is replacing or enhancing the workforce

1031
03:01:28.550 –> 03:01:37.089
Ken Fine: or the workforce model workforce driven model for investing, how that relates to the importance of scalable compute infrastructure.

1032
03:01:38.380 –> 03:01:49.740
Ken Fine: also digging into the roles of agentic AI versus people. And where is it supportive? Where is it replacing? And then, finally, the impact on on founders and operators.

1033
03:01:50.010 –> 03:02:00.000
Ken Fine: A couple of data points to get us started. If you look at investment dollars and percentages in AI and machine learning tools over the last couple of years.

1034
03:02:00.390 –> 03:02:06.580
Ken Fine: In 2024, about 46% of all investing went to AI and machine learning

1035
03:02:06.700 –> 03:02:17.360
Ken Fine: related applications. Whether that’s at the application layer or at the infrastructure layer, you go back just one year. That’s about 36%. So clearly, investment dollars are going in that direction.

1036
03:02:17.470 –> 03:02:30.449
Ken Fine: Similarly, if you look at data center specific investments in 2024, about 50 billion in 2020, about 11 billion. So 2 relevant data points there to tee up the 1st question, which is.

1037
03:02:31.160 –> 03:02:55.740
Ken Fine: as we look at the way investors take a look at and evaluate investing opportunities. How does that analysis or that consideration, balance people and traditional human capital and talent versus scalable infrastructure, Gpu capacity and the ability to scale AI investments

1038
03:02:58.290 –> 03:03:00.230
Ken Fine: who would like to take that first.st

1039
03:03:08.290 –> 03:03:15.349
jeremyk: I guess if no one else is thanks, I’ll start, I mean as a broad, broad set of topics, I mean.

1040
03:03:15.450 –> 03:03:20.540
jeremyk: I think the number one trend everybody’s trying to work their way through is just this idea of

1041
03:03:21.070 –> 03:03:31.870
jeremyk: of, you know, software, the AI actually eating into the labor budget, like everyone frames it as Sas was workflow. And now AI is eating to labor budgets, and I think another way to

1042
03:03:32.130 –> 03:03:44.400
jeremyk: to to talk about the title of this panel kind of Gpu. How did Gpu, how does that translate into labor? So I think one thing that I’ve been trying to work through in my own mind is.

1043
03:03:44.400 –> 03:04:06.900
jeremyk: you know, 2 years ago, 3 years ago, when Chatgpt came out, we tried to just actually do an assessment of like in the different business verticals, whether it’s sales or legal or medical like, actually count the different tasks being done, and then try to think about the number of people completing that task. And then the hourly wage rate. Because I think in this new world we’re all trying to

1044
03:04:07.150 –> 03:04:34.839
jeremyk: think about new opportunities for software to eat work. And it’s helpful to actually start by defining what is the work being done? So one of the things that we did is we counted the number of lawyers, wage rate number of paralegals wage rate. And we’re trying to do that across a number of these sectors. So I think at a high level, we’re all just trying to think about what’s the marginal dollars that eating this labor can capture? Above and beyond workflow? Tam.

1045
03:04:35.920 –> 03:04:46.690
Ken Fine: And Jeremy. What are some of the insights that you found in terms of where the the greater penetration of AI might be? And are they? Are there any bigger patterns?

1046
03:04:46.970 –> 03:05:07.740
jeremyk: Nicely. It coincides very well with the prices venture capitalists are paying for the late stage startups in those respective spaces. So when you do that relative sizing number one and number 2, not shockingly, are code generation and customer support. And there’s logic to why a bunch of the late stage Vcs are paying those.

1047
03:05:07.740 –> 03:05:30.750
jeremyk: you know, very high prices for a 7 million dollars arr business. Just because when you do the relative market sizing, there’s a lot of dollars in those spaces. So I think the 2 that stood out just from a raw dollars horizontal standpoint were code, generation and customer support. And then, in the vertical space, you know. Obviously, you think about a couple of interesting data points. I mean, I think

1048
03:05:30.830 –> 03:05:56.599
jeremyk: you know, I think doctors, financial services, professionals, and lawyers are the ones with the highest wage rates. And then, when you think about what venture capitalists have funded, I mean, it’s not surprising, you know, Harvey Casetext, all the legal AI companies, the medical AI companies. We did that analysis. And then you’re like shoot. Maybe I should have done that 6 months earlier, because it actually has coincided with what a lot of the Vcs have funded.

1049
03:05:59.390 –> 03:06:07.079
Hadley: Yeah, maybe. Oh, go ahead, Thomas. Oh, yeah, I was just gonna kinda add on to what Jeremy was saying. I I think

1050
03:06:07.210 –> 03:06:30.269
Hadley: if you look at kind of the the types of things that are being replaced probably like a couple of years ago. Kind of the 1st things to go were, were the ones that Jeremy mentioned. Because those are the biggest opportunities. And I’m a seed stage investor. A lot of you guys do seed some a little later. I think if you look at where a lot of the new money is going in right now, it’s more vertical specific. I think there’s there’s a notion that that’s what’s more defensible.

1051
03:06:30.536 –> 03:06:38.519
Hadley: And hasn’t been kind of gobbled up by like the 1st wave even kind of your your Harvey’s and whatnot, and at least for us. Yeah, we’re we’re kind of

1052
03:06:38.580 –> 03:06:49.793
Hadley: right. You see, a lot of kind of very niche verticals that you know, you couldn’t build. Software. And before, because that they were, there was so unstructured that maybe if it didn’t make sense to

1053
03:06:50.813 –> 03:07:03.476
Hadley: to build software, and because they were relatively small. But then, when you add the labor to the software opportunity, they they get kind of intrabackable. And then you have the really big ones, and probably where we’ve been spending most of the time are kind of healthcare.

1054
03:07:04.007 –> 03:07:20.859
Hadley: financial services. We we just have like a lot. If you look kind of labor markets overall, like healthcare is the largest sector of the in it. A lot of it is back office kind of manual processes that can be automated. So I think you’re seeing a ton of companies built there that that and and some big ones will emerge.

1055
03:07:22.150 –> 03:07:37.849
Thomas Cuvelier: Yeah, I think I think, what’s what’s interesting as well is that? You know, all the industries that were maybe not typically buying software, also kind of leapfrogging and then starting to to adopt AI tooling. And that’s that’s pretty new. So if you think about

1056
03:07:37.850 –> 03:07:52.980
Thomas Cuvelier: there are even like sectors like agriculture, manufacturing, supply, chain, etc. They were like where software, traditional software penetration was relatively low. Those industries are adopting AI faster than before. So actually.

1057
03:07:52.980 –> 03:08:16.950
Thomas Cuvelier: the time that we see is much larger than it looks. And actually, if you look beyond the obvious industries, there’s actually a much bigger opportunity out there. So I think that’s what especially interesting about AI, I think, is the adoption across all those sectors, and it feels like easy to adopt to a lot of industries.

1058
03:08:19.550 –> 03:08:29.050
Kathleen Estreich: I I think the big opportunities. I would agree with Hadley, I think, on the earliest stage. It’s we’re seeing a ton of vertical very

1059
03:08:29.230 –> 03:08:38.829
Kathleen Estreich: verticals with like very specific workflows for those verticals as kind of big opportunities. And then I think, the the sort of

1060
03:08:38.830 –> 03:09:01.599
Kathleen Estreich: native AI Ux designed products that I think are sort of the the AI native version of something I think, is is like a super exciting opportunity. In the application layer where you kind of rethink like a lot of times. It was a human going, and like pushing the buttons. But if it’s a machine, then what does the new ux actually need to look like

1061
03:09:01.958 –> 03:09:10.200
Kathleen Estreich: and then where does the human orchestrate that from at what level of depth within the application, I think is is pretty interesting.

1062
03:09:12.060 –> 03:09:30.510
Ken Fine: So it sounds like a few common trends and themes. One is vertical specific applications to automate and enhance workflows across those verticals as well as investments it just more generally at the application layer

1063
03:09:30.770 –> 03:09:37.370
Ken Fine: and building AI native applications as opposed to bolting on AI capabilities to existing applications.

1064
03:09:37.520 –> 03:09:58.710
Ken Fine: Are you seeing any trends in terms of investments at the application layer, which is where a number of you invest versus the infrastructure layer which ostensibly could be more horizontal and support multiple, you know, companies supporting multiple different verticals, and how you see your firms spending, investing their dollars.

1065
03:10:00.170 –> 03:10:15.513
Hadley: Yeah, I can maybe start off just because I I historically do both. I think if you look back a year or 2, I’ve kind of moved from more infrastructure to tooling and then now much more application layer. So I’m spending almost all my time at the application layer

1066
03:10:15.980 –> 03:10:40.760
Hadley: and and part of that is because of a seed investor like, I think, investing at some of those lower levels at this point in the cycles is really difficult. We, you know, we we have small funds, and we’re not kind of doing 100 million dollar seed rounds. But also, I think it kind of talks to the opportunity. That that we see again within these verticals. So I think there has been, and I think generally, you see, at least at the early stages. You see that General Trend.

1067
03:10:43.090 –> 03:10:43.840
Ken Fine: Okay.

1068
03:10:45.950 –> 03:10:52.059
Ken Fine: Other comments on that. On that trend application, layer versus infrastructure, layer vertical versus horizontal.

1069
03:10:52.060 –> 03:11:11.890
Thomas Cuvelier: For us, I think historically, we’ve tended to shy away from the infrastructure layer just by, you know, just because it’s it tends to be very capital intensive with with a few winners. There’s always a risk that you can. You can sink a lot of capital and then and back to the wrong company.

1070
03:11:11.890 –> 03:11:25.900
Thomas Cuvelier: So I think in the application layers you can find, like many winners, being infrastructure layer, I think the risk of your commoditization is very real, and the risk of being disrupted is is quite high as well.

1071
03:11:25.950 –> 03:11:27.889
Thomas Cuvelier: And if you think about

1072
03:11:28.000 –> 03:11:46.540
Thomas Cuvelier: the cost or the the capital necessary to to be a competitor in the global stage is is really really high. We’ve seen it in Europe with contenders to Openai, like, you know, the likes of Mistral in France, etc. It’s very costly just to buy an entry ticket in that in that race.

1073
03:11:48.200 –> 03:12:13.429
jeremyk: I would, I would add, I mean, one of the interesting dynamics at the app layer is just then the sheer number of competitors in a defined category. I mean, I I would have thought 4 or 5 years ago. Again I do A and B. So at the A and B stage it was. Usually there were 2 or 3 there, and you could be logical and rational and kind of figure out the one or 2 I mean, I feel like when I’m looking at an A. Now I’ve got

1074
03:12:13.430 –> 03:12:26.889
jeremyk: 8 to to sort through, and somehow you’ve got to sort through them all before somebody else does. And I I’m intrigued at the seed stage. How how you guys are, what you are waiting, when, whether it comes to

1075
03:12:26.890 –> 03:12:38.799
jeremyk: industry, expertise versus technical expertise versus just recognizing. You have to invest in more companies given. You’re you’re in a 1 in 8 world versus a 1 in 4 world, or something like that.

1076
03:12:39.930 –> 03:12:58.479
Kathleen Estreich: I think that’s been the hardest thing at preseason seed, as you see. Used to see. Maybe you know 10 companies. Now, it’s like 20, and they all have revenue. And so revenue is no longer, I think, as as high of a signal, because it’s also understanding that revenue, and where it came from, but I think

1077
03:12:58.520 –> 03:13:21.834
Kathleen Estreich: at the earliest stages to me always it still goes back to the founder like is this founder amazing? And then they’ll kind of figure it out. But I think, weighing that, especially in the vertical side, it’s like the AI expertise plus the domain expertise is sort of that sweet spot of founders for me, where you have the business sense and insights, and then the very strong technical leadership to go

1078
03:13:22.150 –> 03:13:49.019
Kathleen Estreich: build something there, but it is kind of wild. And you know, we run Parex, which is our early stage accelerator, and so next week is my week to go through all the b 2 p. Application applications, and I can only imagine that there will be, you know, 20 plus in in each category of trying to sort through. But for for me as an investor at the earliest stages it usually just comes down to like the best founders and broadly

1079
03:13:49.020 –> 03:13:52.690
Kathleen Estreich: building in a big enough market, but I think the best founders win every time.

1080
03:13:54.080 –> 03:14:10.090
Hadley: I I totally agree with that. And definitely, that’s that’s our approach. I’d say, in addition to kind of grabbing the best founders, and especially like an up like really looking for kind of speed at this point, like the speed of execution, speed of iteration, maybe even more than

1081
03:14:10.210 –> 03:14:24.769
Hadley: in the past. That’s always been important, but especially now we’re also, I feel like we’re kind of all going down towards a longer tail of of like sectors that you can build a vertical solution in and trying to be kind of a 1st mover with a really fast moving

1082
03:14:24.960 –> 03:14:42.550
Hadley: team in a space where you can still build because you’re you’re you’re you’re replacing labor, you a multi-billion dollar company so a little bit that as opposed to hopefully, not having to go into a space where there’s 11

1083
03:14:42.910 –> 03:14:47.160
Hadley: series, a companies, because that that is a seed investor that seems like a tough bet.

1084
03:14:48.520 –> 03:14:58.649
Ken Fine: So given, the heightened level of competition, number of companies being funded and pursuing different applications. Crowded competition.

1085
03:14:59.170 –> 03:15:14.480
Ken Fine: What advice would you give an operator founder CEO, head of product on trying to develop competitive advantage. Differentiation within AI. What are some of the trends that you think and strategies that you see working.

1086
03:15:19.050 –> 03:15:47.950
Thomas Cuvelier: For me. I think the cost of building software is going to strength you to 0. So it’s I think it’s it’s hard to to gain any kind of advantage on the speed to market or so. So in the end it comes down to product. And and then, if you want to build a better product, it comes down to hiring the the best people for so for me, I always advise founders, you know, just hire the best people, and that’s going to be your competitive advantage.

1087
03:15:48.000 –> 03:15:51.950
Thomas Cuvelier: And the second thing is like, be very

1088
03:15:52.050 –> 03:16:17.640
Thomas Cuvelier: thoughtful about how you build the product, because it’s not going to be. You can’t differentiate on the exact value proposition, but you can differentiate through the workflows and the customer experience. I think today the customer experience. The bar is raising significantly, and people just expect more from from products. So for me, I always tell founders, you know. Go back to the basics, hire great people focus on the product.

1089
03:16:17.640 –> 03:16:22.170
Thomas Cuvelier: And then, you know, I think distribution can can sort itself out.

1090
03:16:23.420 –> 03:16:31.799
jeremyk: I think small tweaks in distribution in the early days can make pretty big differences. I mean, I think.

1091
03:16:32.010 –> 03:17:01.579
jeremyk: Hadley, you wrote up healthcare, I mean, I think abridge is the canonical example of that where that that epic partnership really changed the game. I mean, I’m in one of the plg centric scribes, where just the act of letting a doctor try it out. Self adopt, which is, which is highly unusual for that area. I think that’s really given some energy to a crop of companies that might have otherwise faced some hurdles. So just like cleverness

1092
03:17:01.580 –> 03:17:14.820
jeremyk: in early go to market seems to matter. And yeah to your point about what the competitiveness like. I was on Tuesday I was at Clock, which is a legal tech conference in in Vegas, and you just walk in the hall, and

1093
03:17:15.010 –> 03:17:30.599
jeremyk: you know, good for Harvey. Good for like my instinct was, as Hadley said, when you have to go down to a more deeper Ver niche or more verticalized version of that. Because I I was thinking after that conference, like.

1094
03:17:30.640 –> 03:17:55.490
jeremyk: you know, I’m not funding anybody else to go after the Amlaw 100. But I would be interested in funding somebody going after a particular subset of law. But yes, it was quite the experience to walk around a legal tech conference and have 3 generations of legal tech products and 50 vendors, all selling a lot of overlapping products. So I think, as a founder, having that experience would probably be meaningful.

1095
03:17:56.580 –> 03:18:13.010
Hadley: 1. 1 other thing to add on this is, you know, we’re hoping to. We’re we’re in terms of advice I’d give us to kind of have a a second or 3rd act in mind like these things can be hard to predict. But generally I’ll give you kind of an example, there’s a lot of these companies, and and we’ve invested a couple of there these kind of

1096
03:18:13.435 –> 03:18:37.684
Hadley: voice agents for kind of answering the phone, which it adds a ton of value right away for certain types of businesses. But it’s not defensible. So like, okay, you can start with that. And you can actually grow like some of these companies are going from like one to 10 in in one year. But that’s gonna get commoditized. What’s the second act there that you have. Maybe it’s a system of record, or my personal favorite, is it? Maybe there’s a b 2 p marketplace.

1097
03:18:38.120 –> 03:18:41.390
Hadley: which I still think becomes is, is different.

1098
03:18:44.100 –> 03:18:47.000
Hadley: is defensible even in kind of an AI world.

1099
03:18:48.760 –> 03:18:51.640
Kathleen Estreich: I think the 2 big things that that

1100
03:18:51.790 –> 03:19:06.870
Kathleen Estreich: can help a company in the early stages. One is, I, I actually think, like brand and storytelling, and like having some unique reason of or like insight. And then leaning into that, I think brand is gonna become a pretty important part of

1101
03:19:06.870 –> 03:19:24.870
Kathleen Estreich: the way that we, the way that the winners sort of emerge. And and having that DNA early, I think, is really important. And the second is, I actually think distribution matters a ton early distribution. How do you hack early distribution and have some sort of unfair way that you’re going to get your 1st handful or

1102
03:19:24.870 –> 03:19:48.980
Kathleen Estreich: 100 customers. And then how does that create some sort of, you know, unique moat around distribution for you? I think those are 2 things where you know, a lot of the products are actually becoming quite similar. But that’s where. I ask a lot of questions and diligence around those 2 things, and like, do the founders have good answers there, and and good instincts. And maybe it’s my background as an operator on the go to market side. But, like.

1103
03:19:48.980 –> 03:20:16.140
Kathleen Estreich: I just don’t think the best product always wins. I think you also have to win on like the best product, plus differentiated. Go to market, and some sort of network effect, or like unfair partnership or unfair, like founder brand that will help you grow faster is you? Look at every winning company in the tech space. They had something like that. And I think it’s not only like a nice to have. I think it’s mission critical in this like super competitive landscape that we’re operating in.

1104
03:20:16.580 –> 03:20:32.680
Hadley: I have just one more to add, and I and honestly, I hate this. But, like you, you are seeing a decent amount of like King making just raising a lot of money, and then everyone else is scared shitless. I think you saw this with Harvey. I doubt that was the best product from what I’ve heard, but I think there’s a lot of that going on as well which I dislike. But it’s happening.

1105
03:20:33.130 –> 03:20:47.340
jeremyk: I keep on hearing people use the word the fu round. It means like, I don’t need to raise this round. The milestone hasn’t been achieved, but because I can let me detonate that bomb, and then, you know, I’ll make a statement.

1106
03:20:47.660 –> 03:20:50.050
Ken Fine: Make a statement to the market. Stay away.

1107
03:20:50.170 –> 03:21:02.610
Ken Fine: Yeah, this is Myspace. Related question. Let’s say you’re a series later stage series. C series. D, so you’re you’ve been around for a while. You’ve got a business. You’re not AI native.

02:23:43.520 –> 02:23:52.160
Mary Shea: you know, what kind of AI tools are you experimenting with right now? And I’d love to go through everyone and and get a bunch of tools out there.

1108
03:21:02.710 –> 03:21:24.949
Ken Fine: and you’re starting to see competitors come in that that are AI native. So they don’t have the legacy infrastructure and the tech debt and all the ui that you’ve built that was built before AI existed. Any trends or patterns that you’re seeing for how those later stage companies defend themselves and compete and try to prevent the disruption

1109
03:21:25.220 –> 03:21:31.829
Ken Fine: of an AI native entrant like, what’s some of the best work going on for folks who just don’t have the benefit of being born today? They were born

1110
03:21:31.960 –> 03:21:33.010
Ken Fine: yesterday.

1111
03:21:33.900 –> 03:21:49.844
Kathleen Estreich: I I worked at intercom about 10 years ago, and I think that like as soon as you know, like open AI Chat, Gpt came out. They just like pivoted the whole company, and and like went all in on AI support and launched Fin. And I think

1112
03:21:50.170 –> 03:22:04.720
Kathleen Estreich: they like kind of bet the company around being 1st there and and like putting a bunch of resources around that. I think that has been effective for them on the outside. I’m no longer affiliated with the company, but I think

1113
03:22:04.720 –> 03:22:22.530
Kathleen Estreich: to do it well, like you kind of have to do that. Otherwise I think you’ll get left behind. And so whether you take a large percentage of the team to go do that, or, you know, build a tiger team. If you’re not doing it like you’re going to get left behind. I think it’s going to happen sooner than people think.

1114
03:22:22.530 –> 03:22:42.610
Kathleen Estreich: And I think the best companies like I started my career at Facebook, and several times that company had to pick like we were about to go public, and we had no mobile strategy for mobile ads. And Zach pivoted the whole company. I think the best leaders recognize that, and can see around the corners, and aren’t afraid to make that hard fork towards

1115
03:22:42.610 –> 03:22:56.989
Kathleen Estreich: what you need to do, and I think we will see some people hedge and probably get left behind, and the people who make pretty definitive moves there, I think, can stay competitive because you have the relationship with the customer. So it’s kind of yours to lose if you don’t do it.

1116
03:22:57.270 –> 03:22:58.780
Ken Fine: Yeah. Pivot hard.

1117
03:22:59.570 –> 03:23:03.920
Thomas Cuvelier: Yeah, I think if you’re selling to enterprises today, you’re you’re in slightly

1118
03:23:03.990 –> 03:23:19.747
Thomas Cuvelier: better shape, because the you know, the churn will be less. And I think some of those companies will be wary of adopting a AI native company, and probably will will want to wait a little bit, and it’s probably existing relationships or so in place.

1119
03:23:20.090 –> 03:23:43.479
Thomas Cuvelier: but I think if you’re an incumbent selling to small companies where the cost of churning, you know, like the churn is much higher. The cost switching is becoming lower and lower. I think it’s going to be really difficult to innovate. And I think you just have to. It’s going to be a bit of a war to talent, and then you gotta have to kind of like speed up innovation. But I think it’s it’s difficult, right? And that’s

1120
03:23:43.510 –> 03:23:50.160
Thomas Cuvelier: I think that’s what’s so exciting about the new AI companies. They can take market share very, very quickly.

1121
03:23:51.310 –> 03:24:06.999
Hadley: It’s it’s funny, because, traditionally, what you described like a series C Company was the one that I was like most scared of entering, of competing with as like a seat. Someone leading a seed round was like, All right. They’re still like nimble, and but they have resources.

1122
03:24:07.010 –> 03:24:24.629
Hadley: But at least, while there are that have some that have done what what you all are describing. Generally we’re we’re finding them not very strong competition, like they. They’re just too kind of ingrained, with like large departments and and product that’s built, that they have trouble turning the ship and making the necessary

1123
03:24:24.650 –> 03:24:28.300
Hadley: rip, like taking the necessary risks to kind of throw a lot of that away.

1124
03:24:28.587 –> 03:24:33.740
Hadley: To kind of focus on the new World. So at least I think most folks are not doing this well.

1125
03:24:35.150 –> 03:24:43.969
Ken Fine: And I’m speaking as an operator of a series. C company, it. It is hard. I think my experience we’re we’re a system of record

1126
03:24:44.350 –> 03:25:01.430
Ken Fine: is that what we look for, what I look for personally are what are the advantages that we have in data as a system of record so unique sets of data that we have that others don’t. So if you’re a seed based company, you may have great technology. But you don’t have our data set in our case on relationship crafts.

1127
03:25:01.780 –> 03:25:21.359
Ken Fine: The second is vertical expertise, expertise on workflows and just understanding the customer understanding the space and then taking those 2 and saying, All right, what? What are the rifle shot applications that we can develop? We have a right to win, as it relates to AI generally energenic applications more specifically.

1128
03:25:21.360 –> 03:25:38.910
Ken Fine: and try not to get distracted building things that others can build, or you can simply open up your platform, make it more Api based and start to draft off a lot of those other applications that people are building. But pick the subset, that you have a reason to believe that you have a right to win.

1129
03:25:39.210 –> 03:26:02.519
Ken Fine: And then and then, to your comment, Hadley, you do have to take some courage and say, Well, this is our roadmap, and this is our strategy. The world has changed much faster than I think. I’ve seen the world change, and since I’ve been doing this for decades you have to have the courage to make those changes push the company, change strategy, change, roadmap. All that, I think mixes together

1130
03:26:03.210 –> 03:26:06.279
Ken Fine: as a serious, at least for me as a series C operator.

1131
03:26:06.430 –> 03:26:28.179
jeremyk: And I think what’s going to be weird is that we as venture investors, I feel like, if you were to create a a chart like if the rows were the category like customer support. And then the columns were like system of record company. What we call these, like AI teenagers like they got started 2015, 2017, pre chat, gpt, and then, like the newer, the newer post Llm. Companies.

1132
03:26:28.570 –> 03:26:54.620
jeremyk: I don’t know if others share this opinion, but I feel like the end results can be fairly heterogeneous, like there’s nothing structural that says intercom. There’s nothing structurally that says intercom can’t do it. And maybe intercom wins, and I feel like in 5 or 10 years, when we circle the winner in the category by the different rows. I think it’s going to be fairly heterogeneous where, like, there’s no theoretical reason.

1133
03:26:54.700 –> 03:27:06.149
jeremyk: It’s oh, my God! Only the Llm native companies have the way to do it, but it’s gonna come down to choices execution team, which is very odd, because I don’t think there’s

1134
03:27:06.470 –> 03:27:10.489
jeremyk: yeah. I think anyone anyone has the right to win. If they make the right choices.

1135
03:27:10.870 –> 03:27:14.750
Ken Fine: Yeah, the right strategic choices and have the right talent.

1136
03:27:14.750 –> 03:27:28.769
jeremyk: Yeah. Yeah. And so, therefore, you as an adventure investor, I don’t think it’s right to say, oh, you should not do this early stage deal because of this incumbent, because I think in 10 years it’s going to be fairly heterogeneous as to where the winners come from.

1137
03:27:29.360 –> 03:27:45.879
Kathleen Estreich: I think that goes to Hadley’s point earlier. Like velocity matters so much. It’s like teams that can ship. It’s like my favorite thing to ask people between meetings is like what’s new. And if it’s like, nothing’s new, you know, you gotta be moving fast. And I think that slope of building and and

1138
03:27:46.350 –> 03:27:51.130
Kathleen Estreich: speed to execution is is that competitive advantage.

1139
03:27:53.580 –> 03:28:01.235
Ken Fine: I agree. All the years I spent in OP. As an operator and running product. It never seems like we’re shipping fast enough, ever.

1140
03:28:01.690 –> 03:28:04.859
Ken Fine: always trying to, and it’s hard to measure velocity.

1141
03:28:06.820 –> 03:28:10.690
Ken Fine: Other comments on that on that question on that topic.

1142
03:28:14.920 –> 03:28:17.877
Ken Fine: Alright, I’d like to to move to

1143
03:28:18.570 –> 03:28:24.610
Ken Fine: advice. We’ve given some explicitly, some implicitly. Why don’t we start? Let’s say you’re a founder.

1144
03:28:25.060 –> 03:28:32.840
Ken Fine: A seed got an idea you’re prototyping. Let’s say if you’re seed, a you’re obviously further along than that.

1145
03:28:34.310 –> 03:28:36.609
Ken Fine: What advice would you give someone at that stage?

1146
03:28:37.970 –> 03:28:40.559
Ken Fine: Broad question intentionally. Take it where you want.

1147
03:28:49.540 –> 03:28:55.728
Hadley: I can. I can go first.st You want I guess I’ll I’ll take more of the seat stage, because I think

1148
03:28:56.210 –> 03:29:08.000
Hadley: or I guess a as well like. Make sure that you re like iterate with a very small team, and make and do that until you have, like real strong pull from the market.

1149
03:29:08.846 –> 03:29:13.273
Hadley: I talk about this a lot, but like one of the most common failed points is

1150
03:29:13.700 –> 03:29:17.779
Hadley: teams switching from kind of product market fit searching

1151
03:29:18.594 –> 03:29:23.990
Hadley: fit kind of deepening to early scaling

1152
03:29:24.110 –> 03:29:50.340
Hadley: too quickly. And a lot of that ha happens because, founders miss, understand that their their level of true fit. And honestly, a lot of it happens because investors, maybe, who do a new round or whatnot want them to scale, because that’s why they invested and they push them to kind of meet early numbers. So and then they kind of get into like what I’d call like a

1153
03:29:50.965 –> 03:30:04.179
Hadley: a local maxima, where, where, like, you know, they they kind of built the company around a certain kind of product and and go to market and and attempt it just like, Never gonna grow that quickly, and then they can grow just quickly enough to kind of

1154
03:30:04.480 –> 03:30:22.570
Hadley: not, you know, stay alive and look like it’s doing okay, you you know what I mean. So avoid that urge you until like, you’re really getting pulled from the market, and it’s like no one can like you can’t. You can’t service your customers fast enough, and you’re not. You’re not needing to kind of then I don’t think you really have product market fit.

1155
03:30:25.410 –> 03:30:26.140
Ken Fine: Yeah.

1156
03:30:26.310 –> 03:30:33.650
Ken Fine: keep going with responses. I see questions are coming in via chat. So after we get the responses, I’ll pivot to the questions via Chat. But go ahead.

1157
03:30:34.000 –> 03:30:36.950
Thomas Cuvelier: Yeah, for for me. I think my my advice is

1158
03:30:37.090 –> 03:31:00.919
Thomas Cuvelier: to funders is, you know, of course, iterate as much as you can, but try to think about your your long term kind of a moat, because I think today it’s easy to build a you know, kind of a almost a lifestyle business, or like another AI sales agent, or another kind of like automation tool of some kind and and gain some traction. And sometimes you’re pretty impressive. Traction like a couple

1159
03:31:01.090 –> 03:31:17.700
Thomas Cuvelier: of 1 million, and that can feel easy, I think, because it’s just such a big wave of adoption. But I think before raising capital I always encourage founder to think like, what? What are you building? How are you going to make those

1160
03:31:17.700 –> 03:31:36.990
Thomas Cuvelier: those products stick with your customers? And what’s what’s the long term game here, because there’s too many business that I see. That should be, you know, lifestyle businesses, or or maybe the revenue quality is fairly low, with high churn, and so on, and I think that’s something to watch out for.

1161
03:31:37.870 –> 03:31:44.239
Ken Fine: Even if you’re early, what’s your strategic vision? And the moat that goes with that strategic vision? And don’t get too

1162
03:31:44.390 –> 03:31:51.399
Ken Fine: enchanted with some quick uptake of your product. That doesn’t necessarily mean you’ve got a long term path. If you don’t know what that is.

1163
03:31:52.920 –> 03:32:05.390
jeremyk: And also just the fact that I mean things are changing so quickly that I mean, I think in the past you might have been able to formulate a vision, and then over the course of 10 years.

1164
03:32:05.390 –> 03:32:23.310
jeremyk: like, you know, if you saw an Old World Sas product in 2011, you just kept on selling that thing for another 10 years, and the actual product looked much the same. So I mean, today, I mean, just in venture land. A year ago we were all saying, you know, apps are wrappers which I never believed, but like

1165
03:32:23.310 –> 03:32:40.389
jeremyk: it strikes me as hard as a new founder, thinking about it to believe that you can just think through intellectually. Given that there’s so much moving around with Openai. There’s so much moving around with the value of the app layer. So it strikes me as there’s a

1166
03:32:40.390 –> 03:32:51.960
jeremyk: a humbleness that you need where you have to be willing to change, move fast, interact with customers, and I feel like the the grand strategic planning seems like seems very challenging.

1167
03:32:54.140 –> 03:33:18.949
Kathleen Estreich: I think that you need to have some sort of vision of where you’re going, and like, have a set. If you’re going to raise venture money. Then you need to be addressing like a big enough problem where you hit venture scale. I think the way, though, to start at the seed in series A is like tackling something very narrow, and then earning the right to to. Then, you know, expand the vision, and I think sometimes people ship

1168
03:33:18.980 –> 03:33:46.680
Kathleen Estreich: that bigger vision in the v, 1 of their product that is almost like trying to take on too much where you’re trying to be. Everything to everyone versus like solving a very real problem today that then you can expand from there. But I think the challenge there from that I see as like an investor is, you see, a hundred companies like Take Martech, for example. You see, like a hundred companies doing like the you know, the same thing. They’re all going to end up trying to build this like

1169
03:33:46.710 –> 03:34:10.239
Kathleen Estreich: AI marketing team. The question for me always is like, Why is your starting point, the strategic starting point where, whether it is like access to the data or owning, like the most important workflows that then you can build the layers on top of. So know where you’re going. But think really hard about why, that starting point is the strategic starting point that’s going to help you kind of realize, you know.

1170
03:34:10.300 –> 03:34:15.749
Kathleen Estreich: the the layers that you need to build on top of, because if you have the wrong starting point, it’s really hard to back

1171
03:34:15.950 –> 03:34:19.000
Kathleen Estreich: your way into the bigger vision from there.

1172
03:34:19.320 –> 03:34:19.860
Ken Fine: Yeah.

1173
03:34:24.230 –> 03:34:28.749
Ken Fine: agreed. Should we pivot to some of the questions coming in via chat?

1174
03:34:29.350 –> 03:34:31.527
Ken Fine: Julia, have you been running those?

1175
03:34:32.410 –> 03:34:36.110
Ken Fine: Should we just all read them, or would you like me to read it out loud? What’s best.

1176
03:34:36.110 –> 03:34:48.000
Julia Nimchinski: However, you want to take it. Essentially, the question is about founders, and probably the lack of experience. Of some founders. See Hadley has some thoughts on this already.

1177
03:34:49.592 –> 03:34:53.127
Hadley: Actually sorry. I meant to just eat the chat.

1178
03:34:54.260 –> 03:35:04.549
Julia Nimchinski: Oh, yeah, so in short, when you are investing in the founders, how critical is your role as an operator partner in the scaling of the business.

1179
03:35:08.150 –> 03:35:09.030
Julia Nimchinski: Jeremy.

1180
03:35:10.071 –> 03:35:14.350
jeremyk: Sorry. I was just trying to understand the who the the your role, like the you being.

1181
03:35:14.350 –> 03:35:15.370
Julia Nimchinski: Investment Fund. Yep.

1182
03:35:16.606 –> 03:35:22.700
jeremyk: So my role as an operator in helping the company or picking the right company.

1183
03:35:24.760 –> 03:35:42.310
Julia Nimchinski: Yeah. Some of the most innovative, I guess. Agentic startups. The hottest startups are just the founders running them. They just lack experience. So when you invest in those, what is your role as an investment firm.

1184
03:35:42.810 –> 03:35:58.066
Ken Fine: So basically, we’ve got some very inexperienced founders, albeit great ideas. What role do you play as investors to support them and enable them to execute and scale understanding that they, you know, don’t probably have the pattern recognition yet to

1185
03:35:59.150 –> 03:36:00.559
Ken Fine: get out of themselves quickly.

1186
03:36:02.070 –> 03:36:04.529
jeremyk: Sure like I can start. I mean, I think

1187
03:36:04.770 –> 03:36:17.119
jeremyk: I think across the venture ecosystem I mean different stages of folks, whether you’re a precede, or a seed fund, or you’re more of a series, A or series B fund. You tend to help those entrepreneurs in different ways. I mean.

1188
03:36:17.120 –> 03:36:38.310
jeremyk: you know I admire the seed funds that have those databases of like, wow! We’ve literally got a database of like every engineer in the ecosystem, or like every go to market hire. And you know, I think signal fire is an example of a firm that has a database like that. I’m just like, wow! If I was a founder like that’s really cool, that that these seed firms have that information. They can make those

1189
03:36:38.310 –> 03:36:49.659
jeremyk: introductions. I think that’s the value. In the early days. It’s the help me find a great engineer, or you know, I know who the good folks in sales and marketing are, and then, you know, at Series A and series. B,

1190
03:36:50.000 –> 03:37:04.199
jeremyk: you know, we, you know, if we’re doing our job well, it’s helping the companies land the right executives to lead the functions. So that’s something that our firm has developed. But you know, you’ve got 4 investors. I’m sure everyone has a different perspective here.

1191
03:37:08.540 –> 03:37:14.629
Hadley: Maybe one thing to add, kind of, because it’s on this topic. It’s interesting how you go through kind of cycles like, I, I started investing

1192
03:37:14.770 –> 03:37:30.500
Hadley: or 15 years ago, and you had a lot of young founders in the early days of Mobile because it was new. And when you have a new cycle. It kind of gets rid of some of the some of the value of experience so you had your kind of snaps and a lot of these consumer companies are started by very young

1193
03:37:30.500 –> 03:37:47.860
Hadley: people, and then that kind of as we kind of that got more there. There wasn’t kind of a a platform shift. It kind of shifted more. And you saw, like kind of older, more experienced founders, generally getting funded, I think, with Gen. AI and agents. You’ve seen it kind of come back where you’re seeing more and more young founders

1194
03:37:47.910 –> 03:37:52.160
Hadley: with. We haven’t had a job before like creating really cool stuff. And

1195
03:37:52.491 –> 03:38:13.369
Hadley: so yeah, just maybe I’m not sure I’m answering the question. But it is kind of an interesting dynamic. And I think you’re going to see this a lot more, and and it puts more of a burden on investors to kind of work them, and if they don’t know the basics of hiring and go to market like it’s it’s more on us, and making sure that we’re kind of supporting them in in those endeavors.

1196
03:38:15.170 –> 03:38:31.280
Kathleen Estreich: Yeah, I think a pair we I mean, we do pre seed and seed. And we do a lot of kind of young founders like builder founders. And so I think for us, we we take a few different approaches. One is like we run an accelerator that builds in a lot of this support that kind of like every

1197
03:38:31.280 –> 03:38:55.740
Kathleen Estreich: founder needs, and in going from pre seed to seed. So things like hiring like early go to market like, how do you do? Early? Kind of founder led sales like all this stuff? That kind of every company is sort of going through. So we try and like programmatize that to some extent to like help them, and also provide that that peer network of people that are all kind of going through that together, because I think investors play one role. I think

1198
03:38:55.780 –> 03:39:12.309
Kathleen Estreich: having peers and companies that are kind of similar stage going through. That is also a helpful thing. So I think you know you try and programmatize it to the extent that you can build that relationship with the founder. But I think it’s also a delicate dance where you know the best founders.

1199
03:39:12.310 –> 03:39:36.619
Kathleen Estreich: you know kind of can figure it out, and your job as an investor is to ask good questions, you know. Help them get to the answer, get the resources they need faster and accelerate some of that learning. But I think it’s delicate. If then, you have to, you know, handhold them too much. Then I think that’s a crutch, and not necessarily setting them up for success. And one area where at pair we specifically have invested quite heavily is around like hiring. I think

1200
03:39:36.620 –> 03:40:00.579
Kathleen Estreich: that is something where, like, I tell founders, like every founder needs to know how to hire. It’s kind of the same skill set as like going to raise money like you have to convince people to come on this like crazy journey with you. Our talent team. We have, like 4 pretty senior in-house recruiters that help our founders. A lot of it is not just fishing for them. It’s like teaching them how to fish, because that’s going to be their job. I mean, Ken, I’m sure you’re spending a huge amount of your time hiring even at like.

1201
03:40:00.580 –> 03:40:23.820
Kathleen Estreich: you know the stage where affinity is, where it’s kind of the job of the founder always to to do that. Well, and I think I’ve noticed the best companies tend to have these like really high density talent, you know, magnets of companies. And so I think that’s an area where we see pretty high leverage at pair of like helping companies. Kind of figure that out and crack hiring early. And I think that compounds over time for the best companies.

1202
03:40:25.210 –> 03:40:55.179
Ken Fine: Those are great comments, but taking it from an operator perspective both now as a I guess, a more senior operator. But going back to when I was one of the folks who was just learning and straight out of business school. I think the people who are most successful, not just early in career. But I still do this today are those that obsessively seek out, build, develop an expert network. So when I come to a company, for example, I’ve got lots of people who I consider to be best in class in best in class Cros, Ccos Cmos, a level down

1203
03:40:55.380 –> 03:41:19.560
Ken Fine: and pair them with people on my teams and make those introductions to other people in my network who may be early stage founders. So just being relentless about trying to learn from other people who have already done what you’re trying to do, or some portion of what you’re trying to do. So you can learn on their dime for mistakes that they’ve made. The second is you mentioned hiring Kathleen. When I was earlier in career I was

1204
03:41:19.750 –> 03:41:32.509
Ken Fine: always reflecting on my gaps and trying to hire people who had done also what I had not done. So I was usually hiring people almost always, who are more senior than I was, who are more experienced than I was, or at least more experienced me. Technically they may have reported to me.

1205
03:41:32.530 –> 03:41:51.499
Ken Fine: and between hiring and advising, that’s how I think that that can be the difference between scaling and not scaling, because by trying to figure this out, no matter how smart you are, purely through 1st principles and solving problems that way. The rate things change.

1206
03:41:51.990 –> 03:41:55.470
Ken Fine: That’s very hard to do, no matter how gifted you are

1207
03:41:58.220 –> 03:42:00.969
Ken Fine: other insights on that topic. How you.

1208
03:42:00.970 –> 03:42:04.074
Thomas Cuvelier: Yeah, I think for us, we we’re trying to

1209
03:42:04.590 –> 03:42:07.409
Thomas Cuvelier: to kind of to to

1210
03:42:08.060 –> 03:42:24.500
Thomas Cuvelier: to tell founders to build for the long term, because it’s a pretty tricky environment. Because if you’re a founder in AI, and you’re scaling very quickly, you know, you can get away with raising at very high valuations. Money is easy, terms are going to be very favorable.

1211
03:42:24.900 –> 03:42:46.109
Thomas Cuvelier: Often, you know. We see term sheets with little oversight, sometimes like no board seats, and, you know, extremely founder friendly terms, often with secondaries at very early stages. So when you found it, it’s easy to optimize for the for the short term and feel your I mean, you’re feeling on top of the world, especially when you’re scaling

1212
03:42:46.110 –> 03:42:59.109
Thomas Cuvelier: very quickly in the early days. So we we try to, you know. Tell founders to build for the long term and think about what’s next. If you’re raising at a very high valuation, what does that imply for the future rounds.

1213
03:42:59.270 –> 03:43:20.969
Thomas Cuvelier: and also we try to pair them with with founders who’ve seen, you know, seen like difficulties and founders who’ve gone through several rounds of funding, etc. Because at some point, you know, things are going to become more more difficult. But that’s a that’s an interesting trend to see very young founders receiving.

1214
03:43:20.990 –> 03:43:29.300
Thomas Cuvelier: you know, very attractive terms, and that’s often, you know, a recipe for disaster if it’s not handled properly.

1215
03:43:29.960 –> 03:43:42.600
Ken Fine: It definitely can be a blessing and a curse. And you’ve got that high valuation which way I think of it, almost by definition. You need to grow into it. And at any point, if you fall off that track where you’re not demonstrating the financial results that

1216
03:43:42.710 –> 03:43:49.879
Ken Fine: support that valuation that can get very difficult, you know, to raise your next round and to retain people and continue to grow. That business.

1217
03:43:50.220 –> 03:44:02.160
Ken Fine: sometimes going after the highest valuation is not always the best thing. Speaking as an operator, I’ve definitely turned down capital. That was at a higher valuation, because I didn’t think that was best for the business and and the right partner

1218
03:44:06.300 –> 03:44:07.670
Ken Fine: other comments here.

1219
03:44:10.990 –> 03:44:15.459
Ken Fine: Julie, anything you want to hit on the the chat questions that we haven’t hit thus far.

1220
03:44:18.040 –> 03:44:22.710
Julia Nimchinski: Yes, we have 10 min, and would love to address the future of investment.

1221
03:44:23.150 –> 03:44:29.499
Julia Nimchinski: How do you see? Generally like any trends? Gpus over headcount.

1222
03:44:30.290 –> 03:44:32.770
Julia Nimchinski: Yeah, what are your thoughts, Kathleen?

1223
03:44:32.980 –> 03:44:34.060
Julia Nimchinski: Start with you.

1224
03:44:34.918 –> 03:44:38.910
Kathleen Estreich: Sure. I mean, I I think that

1225
03:44:39.040 –> 03:45:01.520
Kathleen Estreich: I’m I’m pretty bullish on, you know the application layer I think we’re at. It’s just a really exciting time. I think the world feels a little bit crazy and chaotic at the macro level, but I think where we invest at the earliest stage on pre seed and seed. It’s it’s really exciting, I think. There’s strong willingness, I think, for people to try new things. At the

1226
03:45:01.520 –> 03:45:26.469
Kathleen Estreich: b 2 b sort of software layer. I was just at a conference a week or 2 ago, and you know the the Cmo of Samsara spoke, and she basically said, no one on her team is gonna get a high rating from a performance review unless they’re using AI. I think you’re seeing these top down, mandates from you know the shopify CEO to, you know, like kind of a lot of a lot of folks have follow suit. So

1227
03:45:26.470 –> 03:45:40.070
Kathleen Estreich: so I think there’s strong willingness to engage in a way that you know there’s a there’s a very strong why, now, of 4 companies going after these markets, I think companies are so lean and and small, which I think is, you know.

1228
03:45:40.330 –> 03:45:49.500
Kathleen Estreich: from my opinion, like more fun way to work where it’s not super bloated. If you’re starting a company today, you kind of get to have all the benefits of this strong technology

1229
03:45:49.500 –> 03:46:12.159
Kathleen Estreich: be an AI native company. And I think the way that companies operate is going to be just so interesting to see over the next. You know, 5 to 10 years where smaller teams can have just huge impact on markets. And and I think there is strong willingness for people to try software in a way that kind of reminds me of, like the early mobile days, where it was just like every, you know, things were like blowing up

1230
03:46:12.160 –> 03:46:37.139
Kathleen Estreich: the long term. Durability and defensibility is, I think, the open question and like, what’s really going to shake out here and what isn’t. But I think there’s I’m like, super optimistic in terms of what we see at the earliest stages. And I think there’s so much excitement, particularly someone mentioned this around like younger founders like, I think you know, these are AI digital native people that are, you know, pretty young in their

1231
03:46:37.140 –> 03:47:02.060
Kathleen Estreich: careers, and able to kind of build these things very quickly, and have, you know, strong traction. Get into conversations with big enterprises. So I think it’s a really exciting time to be starting a company. I think the hard job don’t feel bad for Vcs ever. But the hard job is like there’s so many great companies being started. So I think for me, it’s been kind of sorting through, you know, who are the best founders that have the most unique

1232
03:47:02.060 –> 03:47:13.490
Kathleen Estreich: insight. And and I care a lot about, you know, kind of distribution and go to market. And like, do. These founders have both amazing product sense that can build like a really good

1233
03:47:13.490 –> 03:47:21.240
Kathleen Estreich: product? And can they win on? Go to market? Because, I think you know one is not enough. You need both, particularly in a pretty competitive market like we’re in. Now.

1234
03:47:22.240 –> 03:47:23.553
Kathleen Estreich: I’m optimistic.

1235
03:47:25.230 –> 03:47:35.820
Hadley: Yeah, yeah, I mean, the great thing about being a Vc is that Mark Andreessen says we’ll have the last job standing. So it’s good to have like a lot of job security. Which.

1236
03:47:35.820 –> 03:47:38.548
Kathleen Estreich: I was wondering when that would come up.

1237
03:47:41.310 –> 03:47:59.250
Hadley: yeah, you know I I don’t know. I don’t really care. Share his opinion on this, although you know he’s a smarter guy than me, so he probably knows. But I mean, one thing that I kind of think about is, or wonder about is like what venture and the startup ecosystem looks like going forward. You know, if you take some of the dynamics

1238
03:47:59.370 –> 03:48:06.660
Hadley: around just the the needing less, or you know that the more and more leverage you. You’re getting to build products.

1239
03:48:07.194 –> 03:48:34.470
Hadley: You know. I think you could see a world where where starters could raise a lot less money. And I think you’re kind of seeing seeing a little of the the reaction to that where more and more venture capitalists are kind of focusing earlier, a lot of the multi stages are focusing more of their resources down on seed. And that’s for a number of reasons. But I think one kind of long term reason could could be around. Do these companies? We actually need that much, especially at the application layer. Certainly infrastructure.

1240
03:48:34.470 –> 03:48:51.809
Hadley: It’s kind of a different beast. And then you also have kind of a a decent amount of consolidation, with with fewer funds raising bigger and bigger funds, and like, how are they going to deploy those funds. I don’t know how that shakes out, but it’s it’s going to be a really interesting dynamic. I have a feeling that in kind of 5 to 7 years.

1241
03:48:51.830 –> 03:49:02.160
Hadley: and people have said in this past, and we’ve all been wrong and venture tends to look very similar as it has in the past. I I do think there’s a good chance that venture could look quite a bit different than it than it does now, but but we’ll see.

1242
03:49:03.560 –> 03:49:29.609
jeremyk: I mean just I mean, just even at a personal level. I’ve found that the way I’ve been working and performing diligence in the last year is very different. I mean, I’m you know, looking at a fairly niche opportunity in a vertical I’m not super familiar with like in the Old world. You would go to your associate, and you try to build a market sizing. And now I just run a whole bunch of deep research queries. And

1243
03:49:29.650 –> 03:49:48.110
jeremyk: you know you, you know, as a venture firm, I mean, I think, for the last couple of years a whole lot of the larger funds scale being one of them, but everybody has been doing the scoring like. Look at the the founder backgrounds. Look at the web traffic, and can you create a score? Putting that all together.

1244
03:49:48.110 –> 03:50:10.539
jeremyk: I think that’s just gotten much better. Even quite recently, I mean, with what you can do with Exa functionality, where you can now almost create attribute level groupings on founders. So you can create these unstructured data sets. And then you can almost ask Exa to create the attributes. And I’ve even done things as crazy as

1245
03:50:10.540 –> 03:50:19.209
jeremyk: I literally put in the names of the companies in Chat Gpt. And you ask, you know which one do you think you should invest in, and

1246
03:50:19.240 –> 03:50:36.089
jeremyk: it actually gives the venture consensus answer, which is bizarre, that I mean. I don’t know what it says about venture consensus, but I’ve just been watching my own behavior in the last 3 months on with deep research, with exo, like functionality, and even just chatting with the bot. And it’s

1247
03:50:36.190 –> 03:50:54.130
jeremyk: I mean, we used to vote as a committee. I mean, we vote as a committee. Everybody at scale votes on a scale of one to 10 as an investment. I’m talking to our tech team. I’m saying, why shouldn’t deep research be voting like I mean, I’m not saying it’s right or it’s wrong. But I’m just open to newer ideas on how to think through it.

1248
03:50:55.330 –> 03:50:56.130
Ken Fine: Fascinating.

1249
03:50:59.020 –> 03:51:04.665
Ken Fine: How much of the vote does deep research get get 51%.

1250
03:51:05.100 –> 03:51:07.880
jeremyk: Prove it’s worth. So who knows?

1251
03:51:08.300 –> 03:51:30.200
Kathleen Estreich: On that front. We use a lot of AI tools just in our like in our day to day, of how we run pear. And it’s it’s so nice. It’s like amazing. We have, you know, like AI does the 1st version of our memo based on the deck that we get sent, and we have, you know, scoring for founders based on their background. So it’s like, it’s awesome. I couldn’t imagine having to do this job without it.

1252
03:51:31.350 –> 03:51:37.839
Hadley: Yeah, I I don’t think any junior here has written a a mark, a market section of a memo in in the last few months.

1253
03:51:40.620 –> 03:51:47.579
Kathleen Estreich: We actually had to split our memo out to be like the facts. And then what is your actual opinion? That’s not written by AI.

1254
03:51:48.900 –> 03:51:55.797
jeremyk: Yeah, I mean, it’s gotten to the point where they when an associate actually does one. I I like, do I trust them? Or do I check them with deep research?

1255
03:52:00.050 –> 03:52:07.270
Ken Fine: Alright any comments. Before I start. I was gonna do some summary here. And start to bring us to a close. Julie, is that work.

1256
03:52:08.420 –> 03:52:13.410
Ken Fine: Alright. So real time. I I’ve captured a few points. I’m gonna go with 5 of them.

1257
03:52:14.150 –> 03:52:37.020
Ken Fine: See how well this characterizes the conversation today. So the 1st relates to the importance of go to market fit. I have a philosophy. I use it at my companies that I call nail. It scale it. It’s it’s a bit oversimplified. But the basic idea is that most issues many of the issues that you often face as a operator are when you scale prematurely.

1258
03:52:37.090 –> 03:53:05.269
Ken Fine: and you start doing a lot of money at go to market a lot of money, whether it’s marketing sales, hiring aes or your roadmap before you actually have a reason to believe the product market fit, and also some level of go to market fit. So you’ve got a product that delivers on the promise and can be successful. You’ve got a value proposition that sticks and a model for distributing to the market. Once you’ve got that scale violently, I say so. Nail carefully, scale aggressively.

1259
03:53:05.550 –> 03:53:25.909
Ken Fine: 2. As you’re doing that velocity matters as you start to scale. I mean, it’s speed and agility matters in the nail. It. Phase and velocity matters in the scale, it phase. And that’s with respect to product and to your multiple points. Kathleen, don’t forget about the importance of advantage to go to market.

1260
03:53:25.940 –> 03:53:40.330
Ken Fine: I’ve experienced situations where very smart go to market decisions, boxed out whole groups of competitors to the point where the challenge became becoming complacent because the competitors couldn’t get in, and I had to push herself.

1261
03:53:40.800 –> 03:54:05.120
Ken Fine: And then my 4th is, even if you’re later stage. So you’re not being born or founded today in an AI native world. You need to be bold, pivot, hard, and likely go all in on these these new capabilities, and then 5. Regardless of all of that, it’s still my view, and I think, supported here. It’s the best teams, the best founders, the best people

1262
03:54:05.230 –> 03:54:20.670
Ken Fine: ultimately figure out the right answers. And then, as an epilogue, if you’re going for funding check with AI 1st and see how compelling your pitch and your team and your product and your data are.

1263
03:54:21.110 –> 03:54:24.569
Ken Fine: I’ll pause there. We got a minute or so left

1264
03:54:24.840 –> 03:54:30.450
Ken Fine: any additional comments, agreements, disagreements, plus ones from today.

1265
03:54:35.530 –> 03:54:37.710
Ken Fine: going once, going twice.

1266
03:54:39.220 –> 03:54:42.390
Julia Nimchinski: And we have 2 min left.

1267
03:54:42.972 –> 03:54:50.399
Julia Nimchinski: Doesn’t make it just, you know, shameless self promotion, affinity. We have some customers here on the panel.

1268
03:54:50.900 –> 03:54:51.830
Julia Nimchinski: Let’s go.

1269
03:54:52.320 –> 03:54:53.430
Ken Fine: Oh, for me!

1270
03:54:53.620 –> 03:54:54.250
Julia Nimchinski: Yeah.

1271
03:54:54.660 –> 03:55:01.519
Ken Fine: Yeah, shameless self promotion. So we are series C business. So we’re not AI native.

1272
03:55:01.590 –> 03:55:27.560
Ken Fine: We’re aggressively working on. How do we? Through rifle shot strategy leverage AI for for value. And our strategy is, take the unique data that we have. Take the unique understanding. We have a venture capital venture. Capital workflows, private equity as well, and use that as a way to deliver value that even an AI native company entering our space would struggle to deliver.

1273
03:55:27.910 –> 03:55:47.550
Ken Fine: We started down market with lots of smaller firms, including pair, and then have been moving into larger and larger firms, and then to adjacencies such as private equity and investment banking. So in the form of shameless promotion. If any of you think that might be a fit, please feel free to reach out to me or reach out to anyone on my team.

1274
03:55:48.410 –> 03:55:52.279
Ken Fine: Thank you, Julia, and thank you to the panel and thank you for everyone who

1275
03:55:52.410 –> 03:55:55.449
Ken Fine: participated in was on the call today.

1276
03:55:56.240 –> 03:55:57.799
Julia Nimchinski: Thank you so much. What a pleasure!

1277
03:55:57.800 –> 03:55:58.700
Hadley: Thanks. Everyone.

1278
03:56:00.280 –> 03:56:00.950
Ken Fine: Thank you all.

1279
03:56:01.810 –> 03:56:02.610
Ken Fine: Bye. Now.